cthia wrote:Without even attempting any calculations, which I won't. I can see the inherent problems in my head because I'm a programmer. I know the inherent limitation of computers and algorithms and the limitation of mathematics and the error of mathematics on the error of computers, both taken
accumulatively. Compounded by inexact measurement of distant stellar bodies. Unless you're suggesting a tape measure held from one body to the next to get truly exact measurements vs accepted observable measurements. Small inaccuracies applied to large numbers fed into an algorithm and chewed up time and time again is the true meaning of
GIGO.
I became privy to the problems studying the
percent error of calculators long long ago. My understanding of the inherent problems of the accuracy of algorithms is solid. I was coding algorithms while I was still sucking snot.
The inherent problems are compounded by the inaccurate measurement of stellar bodies,
to the tolerance needed. Using the same man made instruments with computers and the inherent limitations of the algorithms at its base.
When working with large numbers and repeating calculations (as an algorithm does) inaccuracies will add up. The larger the number, the larger the inaccuracy. Advanced computers in the Honorverse will undoubtedly go far to minimize the error, but it is an inherent error that doesn't go away simply because you have a
fast computer. The fact that computers are used as the basis of calculations is the root of the problem.
Percent error calculations are utilized to minimize the problem, but then they too are privy to the same problem. It's like chasing your tail. All roads will end at the
Entscheidungsproblem and
Gödel's Incompleteness theorems and that darn
halting problem.I'm not saying incredible accuracies can't be managed, but the degree of accuracy you are proposing, well, I'm not buying it because I operate under the hood. Everything looks good on paper, until you attempt to program what's on paper into workable algorithms.
You are only considering the mathematics and that it works
on paper, but you are completely oblivious to the applications and the inherent problems in the real world.
That is a common mistake on this forum, discounting the human factor and how things really work in the real world.
You don't truly understand the problem.
I even see related problems inherent in using your NTP-protocol. If to the second is what you want. Uh uh.